Projected lifetime cancer risks from occupational radiation exposure among diagnostic medical radiation workers in South Korea

Medical radiation workers represent the largest group of workers who are occupationally exposed to radiation, and their number is rapidly increasing worldwide [1]. These workers constitute a well-defined population with a certificate of professional qualification, and they include a large proportion of female workers compared with other occupations; this increases the importance of investigating radiation exposure and sex-specific health effects. Although the radiation exposure levels in medical radiation workers are usually low, these workers include high-exposure subgroups, such as those performing fluoroscopically guided procedures, and with long-term, chronic low dose-rate exposures, which may correspond to the type of exposure in the general population. Considering these characteristics, a few cohort studies have been conducted among medical radiation workers to date [2, 3]. Previous epidemiological studies reported an increase in the risk of cancer among workers employed in early years (e.g., 1950s), indicating the need to assess lifetime cancer risk in medical radiation workers based on current radiation exposure levels [3]. Moreover, information on the potential long-term cancer risk considering current exposure levels is helpful to optimize work procedures and develop strategies to prevent excessive occupational radiation exposure.

Risk projection methods can estimate the excess risk of cancer from radiation exposure in populations [1, 4, 5]. These methods have been implemented to assess cancer risk from medical radiation [6–8], radioactive fallout from nuclear weapons testing [9], radon exposure [10], Chernobyl [11] and Fukushima accidents [12]. However, for medical radiation workers, to date, only few studies have estimated lifetime attributable risk (LAR) of occupational exposure in cardiologists in Italy [13] and vascular surgeons and scrub nurses in Korea [14]. In addition, these assessments were based on the simple application of radiation doses to summary tables in the Biological Effects of Ionizing Radiation VII (BEIR VII) report [15]. However, other factors such as organ susceptibility to radiation, sex, age at exposure, and baseline incidence rates significantly affect the risk of developing radiation-related cancers.

The number of medical radiation workers increased significantly in South Korea, from 12,652 in 1996 to 80,115 in 2016 [16]. A historical dose reconstruction was conducted, and organ-specific radiation doses were estimated for all diagnostic medical radiation workers who were enrolled in the National Dosimetry Registry (NDR) from 1996 to 2011 as part of the registry-linkage cohort study [17]. Using these organ-specific doses to estimate cancer risk could provide valuable information of cancer burden from current levels of occupational radiation exposure among medical radiation workers. This approach is particularly beneficial when the follow-up of cohorts is short and cohort members are relatively young.

Therefore, the purpose of this study was to estimate the lifetime cancer risk from occupational radiation exposure and determine differences in cancer risk by sex and occupational group among Korean medical radiation workers. Considering the rapid increase in the number of medical radiation workers and medical radiation practices worldwide, these findings will help provide a timely assessment of cancer risk based on current radiation exposure levels and identify priority groups for radiation protection among medical radiation workers.

This study presents data on potential cancer risks from occupational radiation exposure in a population of diagnostic medical radiation workers based on realistic exposure scenarios. The estimated risk was small in most cases but varied by sex and occupational group. Radiologic technologists had the highest LAR of cancer followed by nurses. Female workers had a higher projected radiation-related excess cancer risk than male workers, although women were exposed to lower radiation doses than men. Both LAR and LFR showed a tendency of temporal decrease with respected to year started work in all occupational groups. The LFR of thyroid and bladder cancer was high for both male and female radiologic technologists. These projections provide a quantification of potential cancer risk based on the current levels of occupational radiation exposure, and they help in identifying priority groups for radiation protection measures among medical radiation workers.

The projected cancer risk in medical radiation workers was lower than that estimated for staff working in cardiac catheterization in Italy [13] and vascular surgeons in Korea [14], although a direct comparison was not possible because of differences in the applied methods and exposure scenarios. A relatively low radiation exposure in our study group (i.e., the median effective dose was 0.78–22.86 mSv, in contrast to 46 mSv [interquartile range 24–64 mSv] for the Italian cardiac catheterization workers and 7.7 mSV among Korean vascular surgeons) would produce a smaller risk. The risk in this group would be even lower considering the tendency of decrease in the occupational dose by year of start of work. However, there was a wide variation in cancer risk by occupational group, and radiologic technologists and nurses who started work in 1991 had a non-negligible cancer risk, particularly female workers. Furthermore, median annual doses were used for estimating the risk in the RadRAT program because of the skewed distribution of radiation doses, and the estimated risk would be higher if mean doses were used. Some medical radiation workers tended not to wear dosimeters regularly [19], and therefore their actual dose levels would be higher than those reported. However, there is also a possibility of overestimation of risk, based on not taking into account possible reductions in occupational dose.

Female workers consistently showed a higher LAR of all radiation-related cancers combined for all occupational groups although these workers were exposed to lower doses of radiation than male workers [17, 27]. This result may be due to the combined effects of the higher radiosensitivity in women than in men (i.e., the higher coefficients of radiation dose in the excess risk model equations), younger age of first exposure in women, and higher baseline incidence rates for thyroid cancer in South Korea [28]. In addition, breast and thyroid cancer mainly contributed to this difference; thus, if both cancers were excluded, the LARs were found to be higher among males. Both the magnitude of baseline cancer risk and the level of radiation dose are important for a full assessment of establishing priorities for radiation protection. Although the level of exposure to radiation is one of the key components of the degree of cancer risk, it is necessary to incorporate the background risk of developing cancers and exposure information when assessing priority groups and designing interventions for radiation workers. These findings suggest that sex is an important determinant of occupational cancer risk, especially among the occupational groups with a high proportion of female workers, including medical radiation workers.

The difference in lifetime attributable cancer risk by occupational group may be attributable to differences in the distribution of age at first exposure and radiation doses between the groups. In this study, radiologic technologists and nurses had a higher cancer risk than other medical radiation workers of both sexes. The age of first employment for radiologic technologists and nurses was younger than that for other groups (e.g., the mean age was 22 years for female radiologic technologists and nurses, and 26 and 29 for radiologists and physicians, respectively), which contribute to the longer exposure duration. Furthermore, radiologic technologists had higher annual radiation doses than other groups and consequently have higher cumulative doses. In addition, the medical radiation workers in South Korea who were exposed to radiation earlier received higher doses of radiation [17, 29], which explains the observed tendency of temporal decrease in LAR considering the years of radiologic practice. This tendency was relatively smaller in radiologic technologists than in other groups, suggesting that excess lifetime risk in radiologic technologists was not decreased significantly by year of start of work, and therefore these professionals may have priority for radiation protection.

Among the cancer sites, the LFR of thyroid and bladder cancer in both men and women was higher than those of other organ sites. For thyroid cancer, high baseline incidence rate may contribute the high LAR and LFR. The thyroid incidence rate, used as the baseline incidence rate in the RadRAT program, was 52.7 per 100,000 for all ages and both sexes combined in 2010 based on the world standard population (http://www.ncc.re.kr), which was significantly higher than that found in other countries [30]. Compared to that of other cancer sites, a relatively high LAR of breast cancer among women and colon and stomach cancers among men may also be related to higher baseline incidence rates. However, the LFR of bladder cancer and leukemia was relatively higher than that of cancers in other organ sites due to the higher coefficients of radiation dose in risk model than other sites [15] and the low baseline incidences in South Korea [28]. However, further studies are warranted to explain this finding. The LAR of a few cancers (i.e., lung, stomach, bladder, and esophagus) was higher in men than in women but the LFR of these cancers was lower, probably because of higher radiation sensitivity in women and a relatively lower baseline incidence in women than in men.

This study has strengths. The first is the inclusion of all diagnostic medical radiation workers in South Korea between 1996 and 2011. Our analysis included all job categories, whereas previous studies limited the cancer risk analysis to specific procedures or occupational groups. In addition, individual organ doses were estimated on the basis of NDR data from 1996 to 2011. In contrast, previous studies obtained exposure information mainly from simulated measurements in short periods. Furthermore, the lifetime risk presented here was based on age- and sex-specific exposure scenarios separately for different occupational groups using data from a previous survey and a government-based centralized dosimetry registry. Our study population was more representative of the general population, and the exposure scenarios presented in this study were more realistic than those reported in previous occupational studies. Moreover, the enhanced projection models were applied to calculate the lifetime cancer risk from occupational radiation exposure.

There are some limitations to address. First, the dosimetry data may have underestimated the real exposure because not all workers wore badge dosimeters regularly [19]. Although KCDC continuously inspects unusual doses and corrects them, we acknowledge this limitation and the need to further evaluate the validity of badge doses. Reconstruction of historical badge doses for those who started work before 1996 is another potential uncertainty for risk estimation that needs to be improved. Second, the lifetime risk estimation of the model had uncertainties related to baseline incidence rates, changes in work practices, and survival rate because these variables may have changed by calendar year. These changes are inherent to risk projection studies. The assumed linear no-threshold risk model in the RadRAT may also be a subject of debate; however, this model was the most widely adopted and best-fitting model for explaining the radiation dose–response relationship based on current epidemiological data [31]. The estimated risks would be higher if no DDREF was applied. Our estimation program did not consider confounding factors or effect modifiers on cancer risk from radiation exposure, and many environmental and genetic factors may have modulated individual vulnerability to carcinogenic effects of radiation. Because the estimated risks reported here were based on scenarios rather than on individual exposures, the risk should not be interpreted as precise individual risk but can inform radiation protection policy at the population level.

This study projected cancer risk from occupational exposure to radiation among medical radiation workers. The estimated risks for specific occupational groups, from highest to lowest, were generally in the order of radiologic technologists, nurses, radiologists, other physicians, dental hygienists, and dentists; however, the estimates for female radiologic technologists were considerably higher and warrant attention. Our findings provide an evidence of the magnitude of potential cancer risk considering the current levels of occupational radiation exposure. Our findings may also be used for establishing priorities for different populations, increasing awareness about the potential risk of radiation procedures, and making comparisons with other occupational and environmental risk factors. Lifetime risk estimation with individual exposure monitoring may help to timely assess the risk of cancer from radiation exposure and provide practical information for protecting medical radiation workers.